![]() I could not find shairplay sources in the internet BTW, even if most probably forked from shairport open sources (illegal if true?).įinally, the hardware also provides a stereo SPDIF interface natively managed by the AR9331 SoC and supporting up to 48 KHz/32-bit sampling clock, which can be connected to an external digital amplifier/DAC (in my opinion SPDIF bypasses the WM8918). I have not tested shairplay through the ALSA interface and I might suspect that the additional CPU cycles of this software layer could conflict with the CPU sometimes required by other sw programs included in openwrt (httpd, ushare, etc.). Besides, shairplay might take almost all CPU when playing from the iPhone and I doubt that it will be able to allow increased resolutions and can efficiently support additional user/kernel mode controls (the high CPU usage that I register is affected by a modification I did on my device, editing airrestart and adding nice -n -13 to shairplay in order to reduce the risk of buffer unnderruns as I also use -b 75 option for increased iPhone promptness on pause/resume controls in any case, even leaving the shairplay parameters unaltered, considerable CPU is used). I also suspect that the shairplay sw player and related sw audio backends are linked to 2/16/44.1 (so, fixed to CD quality, like on shairport open sources). High-resolution files are downsampled to this rate on the iPhone. Anyway, on the iPhone the AirPlay interface streams two channels (stereo) at 16-bit, 44100 Hz in Apple Lossless format. While the WM8918 accepts up to 48-96kHz/24-bit sample rate (if I am not wrong), in general high-end DACs have better characteristics, provide additional channels and many more controls. I noticed that different firmwares are available and some of them are also offering the Linux ALSA driver, which provides a standardized range of control features to many interfaced devices (but maybe only driving external USB audio cards, while the WM8918 DAC might still keep the i2s direct interface). A user-mode driver provides simple write/ioctl system calls to pipe data downstream to the I2S kernel driver, including basic functionalities to control the DAC. WM8918 is natively interfaced to the Qualcomm Atheros AR9331 SoC via I2S digital audio interface. WM8918 includes many controls that I suspect the Music Box will not handle due to hw/sw limitations (e.g., features like analog/digital audio inputs/mics, signal level, volume control, soft mute, parametric equalizer, programmable dynamic range control, anti-clip, anti-pops, additional output, etc.). It can be compared against other DACs by analyzing the respective tech specifications. WM8918 provides analog output to the 3.5” stereo jack socket with ground loop noise rejection feedback and possibly via direct galvanic connection, without inductive components that introduce noise and distortion. I analyzed a bit these low cost devices after purchasing one of them, as they look surprisingly functional, performant and stable even if so cheap and limited in CPU or flash/RAM memory resources. Music Box devices like Soundmate M1-M2/Airmusic NW11/ IB-MP401Air should all be based on similar/same hw and include a stereo DAC that possibly in all cases is the CL Wolfson WM8918 (a chip that can also be found on some mobile phones and on Arduino Yún). What kind of protocol is your NAS using, is it DLNA? You could try the firmware from the Airmusic given in a previous post, but I'm not sure that will help. ![]() You also have nrender, but it does not consume much CPU. But when I use DNLA from my Linux computer I get "nrender" twice taking 35% and 15% of the CPU. You have "/usr/bin/hairtunes" I get "/bin/newshair" when I use airplay on my Ipad. The implementation of shairport seems to be different. Steve you use the firmware of the M1, I use the firmware of Airmusic. I took screenshots of my telnet session if you care to take a look: … 7913648833Īnd, is there anything I can do to try and fix this? I failed to mention I tried several firmwares (1.0.0.5 that came with it and the 1.0.0.2 that some people have good success with). In any case you were right and it seems to be using "Shairport". When streaming with the Iphone I see that "hairtunes" comes up using a fair amount of CPU while when streaming with Qairplay that service does not seem to be up. I did that again this morning and you are correct. ![]() I was able to get into the unit using telnet. ![]()
0 Comments
Leave a Reply. |